With Shizue Izumi (Shiga University, Japan)
Objective: We propose a real data oriented project-based learning (PBL) model of Shiga University. This is for an undergraduate freshmen program to explore data analysis with a background of elementary statistics.
Background: Shiga University will establish the first department of data science in Japan in April 2017. The curriculum consists of a series of courses in statistics and information science, approved by the Minister of Education, Culture, Sports, Science and Technology, Japan. Some courses are taught with ordinary face-to-face lectures and exercises, and others are taught with a proposed model.
Method: Our model is a hybrid course. That is a mixture of online learning using a Japanese MOOC (massive open online course), face-to-face classroom learning, and PBL using real data. All of these teaching techniques are included in a single-term course: fundamentals of statistical inference. For the five week’s online learning component we use materials of Statistics-II course created by the Japan Statistical Society and Biometric Society of Japan. For the fifteen week’s face-to-face classroom learning experience we summarize the course contents and have a group discussion for PBL projects. For PBL we use a real data oriented project which is an exercise of what is known as the Problem-Plan-Data-Analysis-Conclusion (PPDAC) cycle. Real data are derived from public big database including meteorological, health care, traffic, industrial, and GPS. Some examples of PBL projects will be shown at the poster
Results: In 2016, the feasibility of our model was examined through a pilot study in two statistics classrooms of about 80 undergraduate students and about 15 graduate students, independently. Results of class evaluation survey show that the weighted averages for 17 questions in the pilot class were similar to those in previous 3 years when our model was not conducted, and some graduate students commented that they enjoyed their experiences with real data projects and they found them very useful. More students earned a higher course grade in the pilot classes.
Conclusion: Results from the pilot classes may imply that having previous experiences in data analysis among graduate students may help them more appreciate a value of PBL projects. Our model will be repeatedly examined using the ARCS model (Keller, 1987; 2009) for future improvement.